Conversion of non audible murmur to normal speech based on deep convolutional neural networks

Abstract

Communication becomes effective when the speech signal arrives with the profound characteristics. This induced the researchers to develop an automatic system of recognizing the speech signals from the murmurs. Some of the traditional automatic recognition systems are unfit for the silent environments imposing a need for an effective recognition system. Thus, the thesis proposes a method for automatic whisper recognition using the Deep Convolutional Neural Network (DCNN) as the performance of the traditional automatic recognition methods, like Neural Networks, render the unsuitable performance in the presence of the murmurs. The training of the Deep CNN is performed using the proposed Stochastic-Whale Optimization Algorithm (Stochastic-WOA), which is designed by the integration of Stochastic Gradient Descent (SGD) algorithm with Whale Optimization Algorithm (WOA). The input to the classifier is the features that include pitch chroma, spectral centroid, spectral skewness, and Taylor-Amplitude Modulation Spectrogram (Taylor-AMS), which is obtained by combining Taylor series and Amplitude Modulation Spectrogram (AMS) features, of the pre-processed input speech signal. The experimentation of the method is performed using the real database and the analysis proves that the proposed method acquired a maximal accuracy of 0.9723, minimal False Positive Rate (FPR) of 0.0257, and maximal True Positive Rate (TPR) of 0.9981, respectively. Keywords: Murmured Speech Recognition, Deep CNN, Taylor Series, AMS Features, Optimization Technique newline